To be, or not to be variable: comments on variability in autism spectrum disorder

Social communication and interaction are essential to the human condition; some individuals present, however, with profound alterations in these capacities. The most striking example thereof is Autism Spectrum Disorder (ASD), considered to be among the most severe neurodevelopmental disorders in terms of prevalence, morbidity and impact on society.

The first description of autism can be traced to 1943, when Leo Kanner(Kanner, 1968) and Hans Asperger used described behaviours that they separately observed in children, like social withdrawal, desire for sameness, communication/language impairment, stereotyped motor behaviours, and intellectual disability with onset from the first year of life, but also the expression of exceptional isolated talents and conserved linguistic abilities. Early infantile autism has been included in the diagnostic manual DSM-III in 1980, to be later covered by PDD (pervasive developmental disorder) in the DSM-IV (1994), changed into the current umbrella term of Autism Spectrum Disorder (ASD) in DSM-V (2013). ASD is defined as a set of developmental disorders that are highly variable in presentations across individuals but are behaviourally characterised by altered social-communication skills and sensation, and restricted repetitive patterns of behaviours and interests. A diagnosis of ASD according to the DSM-5 criteria requires that symptoms cause clinically significant impairments in social, occupational, or other important areas of functioning.

ASD is one of the most common neurodevelopmental disorders, with a population prevalence estimated to be between 1% and 1.5%. As stated in a recent and controversial article published by Mottron & Bzdok in Molecular Psychiatry, “the evolution of the DSM has been accompanied by a 20-fold increase in the reported prevalence of ASD over the last 30 years, reaching a current prevalence of more than 2% in the United States”. One of the possible explanations to the increase in ASD prevalence over the years is, actually, the changes in diagnostic criteria, which evolved from a narrower definition of autism to a broader definition of autism spectrum. As stated in a previous article from the same group: “Our understanding of autism has evolved from a narrowly defined clinical picture to a spectrum of conditions of uncertain similarity” (Rødgaard et al. 2019). Of note, they rely on the idea of autism as a very separate condition compared to the umbrella term introduced as ASD, and attribute the increase in diagnoses to the introduction of this non-specific term. Nevertheless, the authors did not mention that the DSM-V lists more stringent criteria for ASD and did not consider other concomitant factors that likely play a crucial role in this increase of ASD diagnoses in the past few years. For example, they did not mention changes in reporting practices, policy changes, increased understanding and identification of under-represented populations (e.g., females), increased funding, increased public awareness, increased availability of services, decreasing age at recognition, changes in referral patterns, improved diagnostic processes, wider screening, increase in parental age and other perinatal factors.

To support their claim of a too broad definition of autism as ASD, Mottron and Bzdok highlight how differences between individuals with autism and typical comparison groups have decreased over time. Of course, they attribute the cause of this phenomenon to changes in the definition of autism from a narrowly defined population toward an inclusive and more variable population. This is accompanied by a consistent and significant decrease in effect size of findings in biomedical sciences, which means that researchers find very small differences between groups. Thus, does it make any sense to have such predefined categorical groups? Probably not, but there are different views on the implications. Before proceeding, it is important to highlight that such small but significant differences between groups found in the latest research is accompanied by an increase in the number of people included in these studies (sample size). Just think about large research efforts put into international consortia (ABIDE, EU-AIMS, etc.) to build large datasets and improve our understanding of the mechanisms underlying ASD symptom emergence. The more people are pooled together in these large datasets, the more variability is included in these datasets. This variability can explain the reduced differentiation between ASD and non-ASD groups. The point here comes to: is this variability just noise, or something intrinsically characteristic of autism?

While Mottron and Bzdok see this variability as an artefact introduced by the loose diagnostic boundaries, and interpret the reduced effect size of research findings as a consequential negative effect, we have good reasons to think that this variability is a key characteristic of autism, which is better and better captured by study designs of increasing quality over time. It is known, in fact, that clinical manifestations of autism are highly variable across individuals in onset, cognitive function, language skills, symptoms profiles, severity, and psychiatric and neurological comorbidities. This variability in manifestations reflects the high variability in causative contributions, which include environmental and medical risk factors (e.g. prenatal drug exposure, preterm birth or congenital infection), and hundreds of genetic variants of small and large effects, impacting specific molecular pathways and regulating fundamental processing of early brain development (e.g. cortical organization, excitation/inhibition balance and connectivity), and ultimately leading to distinct clinical presentations. This variability makes it difficult to understand, to detect and to treat ASD as a single disorder at the level of the individual. How do we handle this? Should we just reject the ASD term, burn our umbrella? Well, Mottron and Bzdok suggest to neglect this variability, consider it just noise, and define a narrower category of autism which might be more identifiable, more “prototypical”. Yes, they want to burn the umbrella, cut it down into pieces. Yet, there are various issues with their suggestions. First, what is this “prototypical” autism they talk about? Isn’t it the main issue here that current research can’t define a “prototypical”, coherent autism category? Who is supposed to define it, then? Second, if we define a “prototypical” autism category, do you find it realistic to keep autism as a separate entity, a “very atypical” group of people? This looks like rather reductionist to me, to go back to biomedical essentialism and deny neurodiversity. Rather, autism, like other psychiatric conditions, would be better described as a continuum at an individual level, reaching extreme values in a spectrum of continuously varying traits in the general population rather than as a well-defined, separate category. Major initiatives in psychiatry have started to focus on different approaches to understanding what seems to be natural variability in mental health. As an example, normative modelling is a promising method to understand variability in the general population, used to evaluate deviance from normative ranges at an individual level by mapping the full range of population variation and considering, for instance, psychiatric symptoms in patients as deviations from the normative patterns.

Finally, Mottron and Bzdok have a point in highlighting the necessity to identify possible subgroups in the ASD category, which might be more relevant target groups for specific intervention programs. Yet, their conclusion stating in strong words that current acknowledged variability of ASD is an artefact and not a creation of nature has no empirical basis. The authors have simply not provided evidence to support their claims and ignored relevant scientific findings and other explanatory factors driving conclusions against their claims. Furthermore, rejecting the ASD spectrum idea for variable conditions sharing the same core symptoms to define ad hoc a “prototypical autism”, in other terms an “easier to investigate” category, seems more artefact than nature. It is true that the idea of ASD as a discrete, separate entity per se can distort the investigation of the underlying mechanisms and early development of the condition. Yet, while it seems a logic consequence to abandon the unitary ASD diagnostic label and investigate the complexity behind it, the ASD category is useful on a practical level for providing a diagnostic label, and with it, access to services. Future efforts should focus on the characterization of this variability, to embrace complexity of nature, to accept neurodiversity, rather than simplify it in a reductionist picture of mental health.

References

Mottron, L., & Bzdok, D. (2020). Autism spectrum heterogeneity: fact or artifact? Mol Psychiatry. doi:10.1038/s41380-020-0748-y

Rodgaard, E. M., Jensen, K., Vergnes, J. N., Soulieres, I., & Mottron, L. (2019). Temporal Changes in Effect Sizes of Studies Comparing Individuals With and Without Autism: A Meta-analysis. JAMA Psychiatry. doi:10.1001/jamapsychiatry.2019.1956

Kanner, L., Autistic disturbances of affective contact. Acta Paedopsychiatr, 1968. 35(4): p. 100-36.

Asperger, H., ‘Autistic psychopathy’ in childhood, in Autism and Asperger Syndrome, U. Frith, Editor. 1991, Cambridge University Press: Cambridge. p. 37-92.

Marquand, A.F., et al., Understanding Heterogeneity in Clinical Cohorts Using Normative Models: Beyond Case-Control Studies. Biological Psychiatry, 2016. 80(7): p. 552-561.

Tordjman, S., et al., Gene x Environment interactions in autism spectrum disorders: role of epigenetic mechanisms. Front Psychiatry, 2014. 5: p. 53.

Christensen, D.L., Prevalence and Characteristics of Autism Spectrum Disorder Among Children Aged 8 Years - Autism and Developmental Disabilities Monitoring Network, 11 Sites, United States, 2012 (vol 65, pg 1, 2016). Mmwr-Morbidity and Mortality Weekly Report, 2016. 65(15): p. 404-404.

Baxter, A.J., et al., The epidemiology and global burden of autism spectrum disorders. Psychol Med, 2015. 45(3): p. 601-13